|Frequently asked questions about learning with SuperMemo|
|Do you have a question about memory, learning or SuperMemo? Write to Dr Wozniak|
See also: FAQ: Learning with SuperMemo
(Constantin Ilieu, Bulgaria,
In your materials I found a contradiction. On one hand you claim that once learned knowledge is constantly maintained in the student's memory, on the other you say that after ceasing repetitions, I will gradually forget what I have learnt. Which is true?
Both facts are true. The term maintained is understood as kept in memory by means of repetitions, not as remains in memory for ever
(Elena and Rachel, US, Dec 12, 1997)
How is SuperMemo supposed to improve someone's memory?
We do not claim that SuperMemo improves memory. We say that it allows you to learn fast with high retention of knowledge. The fact is that improving your memory will be a nice side effect, but this will happen not by virtue of the SuperMemo method but by virtue of intense learning. SuperMemo makes it possible to learn fast by organizing your learning process. For more see: General principles of SuperMemo
(Crizeldo G. Cariaso, MD, Philippines, Dec 17, 1997)
I am a medical doctor who is into resident training as of this moment. Do you advice me to use your product if I read about 100 pages of documents and books a day and have to remember it?
Absolutely! Read Devouring knowledge to see how you can read thousands of independent articles at the same time. If you learn from paper books, your job will be much harder. You will only be able to remember a fraction of the material. An exemplary algorithm would be:
The sad fact is that reading 100 pages daily is really a feat, and even the mere typing in the material to SuperMemo will limit you to 20-100 questions per day (depending on time available and the speed of typing). Even this small proportion will still provide you with amazing build-up of knowledge! It is very important that you intelligently select what must and what does not have to be memorized
Prohaska, USA, Oct 29, 1998)
I am going back to medical school after a 10-year break. I look for tools to improve my brain skills. Is SuperMemo a product for me?
That's your lucky day! SuperMemo is ideally suited for enhancing your ability to learn new stuff and make sure it stays in your memory for good. Learning medical sciences with SuperMemo comes second in popularity to learning English! Recommended reading: Six steps to excellent memory
We believe SuperMemo is a must for anyone with serious plans to enter science
(Garry Gross, Saturday, August 25, 2001 12:24 AM)
I am presently studying to be a behaviorist. I am working from text books that I find difficult. Please let me know how SuperMemo would be useful in this adventure. How do I get the material from the text into the program?
This article is probably the best summary of the role SuperMemo could play in your work: Devouring knowledge. Please pay a special attention to the part devoted to incremental reading. We believe that incremental reading is a must for anyone with serious plans of working in science. The most painful limitation in your context is that incremental reading requires your texts to be available in electronic form (e.g. if you rely on paper books, you would need an OCR scanner or you would need to type in the most essential study points into the computer)
Usela, Czech Republic, Nov 2, 1999)
You advise to merge all SuperMemo collections in one. Does it not contradict the principle of avoiding knowledge interference?
No. The same items are not likely to interfere more with each other once they are put in the same collection and repeated one after the other. Indeed, this may even help resolve the interference by means of techniques described in 20 rules of formulating knowledge. For those who learn material related to various branches of knowledge, an interdisciplinary mixture of repetitions often generates unexpected associations that can be a goldmine of ideas in your professional field. This is the simplest illustration of the fact that SuperMemo can make you more creative!
SuperMemo can lead to learning garbage
(anonymous, Jan 20, 2007, 04:59:15)
If you learn garbage with SuperMemo, you will waste time on reviewing garbage, and possibly retain garbage in your memory for many years
It is true that persistence of memories formed by SuperMemo carries a risk. It is then your sole responsibility to make sure you keep meticulous notes on sources and their reliability wherever your mission-critical information is concerned.
Learning garbage may seem like an anti-thesis of good learning. However, rarely do we face neat textbook information that can serve as model knowledge requiring no verification. In real life, we often face chaos of contradictory information coming from different sources. For example, when working in a fast-growing field of research, you will constantly meet new findings that do not fit old models. Reconciling the chaos of new information may then become your primary preoccupation. Cognitive research shows that our predictive capabilities quickly become saturated with the inflow of additional information. For example, a handicapper may need good information on horses and jockeys to set the odds. However, his accuracy does not improve much once his cognitive capacity becomes saturated. The same extends to other fields where human judgment is involved.
SuperMemo can help you shift the saturation point by making it easier to resolve contradiction. In conditions of low retention, an individual facing contradictory propositions A and B will often oscillate between A and B points. Such an oscillation is a function of exposure over time. However, if you memorize the proposition A, encounter with the proposition B is unlikely to produce such an oscillation. Instead, a red flag will be raised signaling the contradiction. You can then rephrase the question on proposition A to include proposition B with sources and other considerations predicating on validity of A and B.
By resolving contradictions, SuperMemo helps you built coherent models of reality. These in turn improve your judgment and your problem solving powers.
Still, you need to remain vigilant. When trying to understand the world, you are bound to face information garbage. Paradoxically, learning garbage can lead to the emergence of improved models of reality. On the way towards the ultimate model though you may need to face false information that is often more costly than ignorance. You cannot let your guard down.
Music and incremental reading
(Mohammed Asad Khan, Pakistan, Thursday, May 02, 2002 1:01 AM)
Can I read articles in SuperMemo with the help of my favorite music?
Optimally, you should have your favorite music turned off when learning. However much it invigorates your brain, it will ALWAYS decrease your focus and take away some mental processing power. Invariably, background music reduces the efficiency of working with SuperMemo. If your learning is boring, you must diagnose the reasons. Most often, user knowledge is not properly formulated for active recall. Incremental reading may require a few months of practice to develop good learning habits. You cannot resolve the "boredom" of learning with background entertainment. Learning must be entertaining on its own!
You can easily learn 10,000 items per year
(Robyn Harte-Bunting , Monday, October 06, 2003 4:39 PM)
In the website you state that it is difficult to maintain a schedule of more than 100 daily repetitions or 1 hour per day on SuperMemo for lengthy periods. Can you say more about this number?
Those rather arbitrary numbers depend on many factors. Beginners should take far lesser loads. Advanced users may go well beyond that, esp. with incremental reading and related skills. Nothing will serve as a better guidance as your own experiments probing the difficulty of knowledge you plan to master, your skills in formulating knowledge and the degree to which your personality or the type of acquired knowledge will make the whole process enjoyable as opposed tiresome. Some anecdotal facts may help you in measuring your maximum load:
As for incremental reading, advanced users report little or no fatigue even in 3-5 hour non-stop runs (assuming a fresh unstressed mind). This may partly come from the material variety (randomized presentation of topics has a "TV channel zapping" reward effect). Also, the passive nature of topic processing may contribute to resolving short-term memory overloads that quickly result in homeostatic fatigue in traditional SuperMemo.
All in all, if you master all skills related to incremental reading and knowledge formulation, if you test your personal persistence and self-discipline, you may be able to commit yourself to 100 items per day plus 100-400 topics per day. However, as this may take 2-4 hours of learning, it is highly recommended you split those high loads into portions. If you give up on retention (say down to 85-90%), you can maximize the learning speed and reach 30,000-50,000 new elements per year. Of these, depending on the strategy, 30-60% will be items (still rather less than 15,000-20,000).
Remember, that you will need to use your best creative hours to sustain this process in the long run. Assuming you sleep in accordance with you natural circadian rhythm, those hours would be early in the morning (after breakfast) and perhaps in the evening peak (if you experience one). If you happen to learn at sub-optimum time (e.g. after work), you may find it hard to do a quarter of the suggested load
On what basis do you ground your claim that SuperMemo increases the speed of learning from 10-50 times?
For knowledge retention of 95%, it can be computed that the number of repetitions in an average learning lifetime (i.e. about 55 years) is roughly 50 times greater for equally spaced repetitions than for progressive repetitions (as used in SuperMemo). For repetitions with no regular spacing scheme, this number may even be greater. Moreover, the greater the required knowledge retention, the greater the increase in the knowledge acquisition rate (classical forms of learning almost never reach knowledge retention above 10%!!!). In practice, users of SuperMemo claim that it increases their speed of learning from 50% to 2000%. These values are, however, highly subjective, as they do not account for so-called intractable items, which are practically not memorizable without SuperMemo. In other words, students tend to underestimate the fact that they reach knowledge retention from 90-99%, which would hardly be achievable using any other method.
(Shaun Hoffland, UK, Oct 22, 1998)
Do you have statistics on what is the average number of items an average user has to repeat per day? What is the greatest number of memorized items?
The best results oscillate around 60,000 items memorized overall, and 10,000 memorized within 4-5 months. Most users memorize 1000-10,000 items per year. Please read SuperMemo User Survey (esp. section Using SuperMemo)
(Jason Roos, USA,
Mar 27, 2001)
When I need to memorize a bunch information, I make flashcards. My gut is usually a good indicator of when I need to go through them again. It's surprising to me there seems to be so much scientific study behind SuperMemo. (full text below)
If you try SuperMemo for some time, you will notice that it is quite liberating to let the computer do the guessing for you. If you later go into thousands of memorized elements, your "gut" will fail you at some point. After all, it is not easy to guess which of your 100,000 items is closest to being forgotten
Jason Roos: When I need to memorize a bunch information, I make flashcards:
- I write a question on one side and an answer on the other
- I gather all the cards up, and go through them one at a time
- Those cards I get wrong, I set aside in a "don't know yet" pile. Those I get right, I set aside in a "know it" pile
- After I've passed through the stack once, I pick up the "don't know yet" pile and repeat this sorting process until they've all been memorized
My gut is usually a good indicator of when I need to go through them again. If I feel confident, I know I'm ready to take the test. Intervals are not that accurately determinable, and 4 days for the first one, in my experience, is waaaay too long [in SuperMemo, the first interval will depend on the user and the learning material]. I need to pass through them every 30 minutes to an hour for at least a day, before they're reasonably secure in my head. Of course, it all depends on the type of material and how many days I slept in class.
This method [SuperMemo] is hardly unique-- that's how many people study. And the effect is identical to your unnecessarily complex and rather tedious, I think, worksheets and software. It's surprising to me there seems to be so much scientific study behind it. I find it fairly simple and obvious
The effectiveness of passive review may be dismal
(Janusz Jezierski, Jul 18, 2002)
If active recall is so important, what is the purpose of topics in SuperMemo? How efficient would learning be if I never used items?
The main purpose of topics is to collect learning material that will be processed to formulate questions and answers. With topics you do your reading. With items you do your learning. Learning based on topics would be dismal in quality. Experiments show that you may review a sentence passively 20 times, and after the interval of just forty days you may be unable to recall you have ever read or seen the sentence! The more material your process in learning, the more likely you are to see this scary weakness of human memory. The power of SuperMemo is in simplicity of questions, active recall of items, and reviewing the material regularly as demanded by SuperMemo
Poland, Apr 15, 2001)
While I make my repetitions, I often have problems with focusing on questions. My brain simply wanders off to other subjects. Do you think this is normal?
This is a clear indication that your material is either badly structured or poorly selected. Unfortunately, this situation affects quite a number of users. Repetitions affected by concentration problems are dramatically less efficient. A wild guess is that you might be losing 50-95% of your efficiency. If you formulate your material poorly, you may suffer from recall problems (answers are too complex) or comprehension problems (questions are too wordy). Repetitions immediately become less enjoyable or even painful. This will quickly trigger natural defense mechanism. Your brain will search for more productive mental effort such as pondering over the movie you had seen a day before. Poorly selected material will produce exactly the same result. Optimally, you should only store top-importance material in SuperMemo. This comes from the fact that usually you hardly have time to master top-importance items, let alone anything else. If your material does not meet this criterion, it will seem less relevant and simply boring. If you cram for uninteresting exams, the result will be the same. Poorly selected or structured material is the chief cause of showing little progress with SuperMemo. Problems with attention are a clear indication that you need to review your learning process carefully. Probably you will need to do lots of deleting and reformulating.
In addition, be sure you get enough sleep, avoid stress (e.g. make repetitions before school or work), do not let anyone disrupt you, turn off the radio and TV, turn off your music (even those pieces you love will compete with your items for your brain power). Please read: SuperMemo Decalog, esp. the section entitled Concentration
SuperMemo discourages mindless memorization
(Stephen R. Diamond, Jul 21, 2004, 17:55:23)
Does SuperMemo encourage excessive reliance on memory, on using memory where reasoning would be more productive? Since there is no real way to determine how much memorization is beneficial, there is the risk that SuperMemo will dictate what is learned. There might be little material that most people must memorize, with learning the vocabulary of a new language the most obvious exception. The student enamored of memorization technology will be spending his time learning details, not acquiring the ability to determine which details are worth attending to.
There is much to remember: There may be little material that people must memorize, but there is a world of material were memorizing things brings benefits that outweigh the cost of memorizing. If the cost of indefinitely retaining a piece of knowledge in memory is 1-4 minutes, the number of pieces of information that could return this investment is greater than lifetime of memorization. Those pieces may range from a useful mathematical formula, through a location of a country on the map, through the name of a carcinogenic dish, to seeming trivia such as … facts about your mother-in-law (Outlook Calendar entry won't let you sneak a compliment at the family table)
Deductive reasoning feeds on knowledge: We have not yet developed a theory of everything or a super-brain that could use it to derive all truths via deduction. Reasoning requires knowledge. Einstein also needed facts and rules developed by his predecessors to arrive at his theories. Even a genius mathematician working in the attic in detachment from reality will need a few things in his memory to start with (such as a problem to solve in the first place, or a few formulas or theories that will shorten his path to the goal). A physician at the scene of an accident must instantly retrieve seemingly mindless facts from his or her memory such as the name of the drug and the dosage. Relying on deductive reasoning and smarts will not do
SuperMemo provides usability feedback: The central power of SuperMemo is in providing you with full control over what you decide to remember. In addition, incremental reading can serve as a storage of things you care about but not necessarily remember. These storages of things remembered and things that are important provide for immediate associative memory for problem solving on one hand, and a tool for "brainstorming with yourself" and extending knowledge via incremental reading on the other. SuperMemo provides you with a feedback on how knowledge impacts your life. With time, it helps you understand what is worth memorizing and what facts are time wasters. Ignorance may initially tempt you to abuse SuperMemo, but SuperMemo will soon demonstrate which efforts are futile and which bring a palpable benefit. This self-correcting feedback mechanism is a priceless added benefit of getting enamored with SuperMemo.
Not only geniuses use SuperMemo: SuperMemo is used by a special lot of people. Primarily they are a strong-willed and self-disciplined group. Weaker souls drop out very soon. Some may show obsessive and compulsive attitudes and these will backfire. Some may persist for months in a strong belief SuperMemo is the right way to go and still make little progress in areas other than word-pair learning. Like in any group, there are stronger and weaker students. There are crammers and high-fliers. In other words, the uncritical trust in SuperMemo can have negative side effects, but the feedback mechanism mentioned earlier acts as a countermeasure. All in all, with its tangibility and measurability, SuperMemo helps bad students understand their errors. With time, users of SuperMemo show remarkable progress in understanding how their memory works and what the role of knowledge in life is. Bad students will either improve or drop out.
All good things can be used in a bad way: Apart from the side effects of misconstrued memorization,
SuperMemo might also be used by a bad lot for their purposes. Hopefully, Osama is too contemptuous of "western" technology to make his Pocket PC repetitions in the cave
to extend his knowledge of the best ways of harming his
hate targets. But even here we might hope that a dose of extra learning would
bring him closer to seeing that what we share is more important than what
All innovations and inventions can be put to bad use. Cars kill and pollute. Airplanes can be flown into buildings. TV is numbing our knowledge selection skills. Web was branded "a garbage bin of human knowledge". Even books encourage passive review as opposed to active processing. SuperMemo does encourage memorization and even the most proficient user will memorize some ballast of junk knowledge. What matters is the balance of costs and benefits. The value of knowledge is hard to measure. The lifetime impact of SuperMemo is hard to measure. To ultimately answer the above question we would need an equivalent of the longitudinal Terman Study. Will those kids who contracted the bug of SuperMemo in the early 1990s get bogged down with lifetime of time-wasting repetitions? Or will they become presidents, Noble winners, and great inventors? Time will tell
(Prof. Witold Abramowicz, Poland,
Does the minimum information principle not stand in conflict with the ages old rule that the learned knowledge should be highly associative in nature?
No. The minimum information principle concerns the representation of knowledge in SuperMemo databases, not in the student's memory, and it does not prevent great advantages coming from proper structuring of the learned material. In the optimum situation, the student should first construct a cohesive model of the learned subject, and only then, apply SuperMemo to make sure that the learned knowledge is sustained in memory as a whole. The knowledge may be highly associative, but strictly targeted neural stimulation, achieved by means of granular representation of knowledge in SuperMemo, is necessary to effectively induce molecular processes responsible for memory formation. Indeed SuperMemo has been conceived in such a was so as to make it easier to formulate knowledge in a structured way (topics) and later learn it in a way typical for SuperMemo (items). See also: Topics vs items
Bahrick's research on Spanish vocabulary retention does not undermine
(Desmond Connor, Jan 28, 2005, 9:38:48)
Isn't the whole concept of SuperMemo a bit shaky in the light of Bahrick's research on Spanish vocabulary (spanning 50 years!). For example: "The analysis yields memory curves which decline exponentially for the first 3-6 years of the retention interval. After that retention remains unchanged for periods of up to 30 years before showing a final decline"
Bahrick's research is invaluable as it is quite unique in the field otherwise dominated by research with short-term goals. His results are consistent with the two-component model of memory underlying SuperMemo. What may be confusing at first is that the purely exponential nature of forgetting comes from a model of unitary traces of memory. In real life, memories are always composed of more complex engrams. Moreover, bodies of knowledge are composed of memories that are heterogeneous in that individual engrams differ in their difficulty, stability, and retrievability. Finally, Bahrick's methodology, for obvious difficulty with long-term research (incl. the size of the sample), inevitably carries a substantial margin of error. What may appear as a kinked forgetting curve can well be approximated with a continuous power function that is a result of the superposition of exponential forgetting curves in full agreement with the two-component model of memory.
The forgetting of a language can best be illustrated with the observation that after a course, we usually quickly forget a large body of knowledge. Usually the most difficult and the least rehearsed knowledge goes first. Once this body is eliminated, the cumulative forgetting curve flattens. Now those easier pieces of knowledge and those that have been best rehearsed decline at slower rate. Their individual forgetting curves are much flatter. Overall forgetting is slower. Finally, many years later, only the most stable memories survive. Those may be entirely resistant to forgetting. This does not mean that these are permanent memories. Most often, these are memories that can be refreshed by association: a face of a friend from school, thoughts of youth, an old movie, or a daily ritual. After all, it is quite hard to forget Hasta la vista baby
I have an exam for a driver's license in 2 weeks. How can I best memorize the Traffic Regulations collection for SuperMemo? How can I increase the frequency of repetitions?
Although SuperMemo is not a cramming tool, and it would be much safer to start 2-3 months before the exam, the following shall work pretty well:
Why is SuperMemo slow to show its strength?
(Nathan Crow, WedJun18,2003 4:14 pm)
Why do you say one cannot see any difference between SuperMemo and other flashcard systems until a week has passed? Odd, if true.
SuperMemo uses spaced repetition. Inter-repetition intervals increase gradually in the learning process. The more you grow your intervals, the less time you spend on repetitions, and the more time you have to enjoy stable knowledge.
If you compare SuperMemo with reading a book, you will see little difference in retention in the first week (first intervals in SuperMemo are usually shorter than a week). If you compare results after a month, an average book reader will have already forgotten a large part of the learned material, while the user of SuperMemo will stabilize around the programmed level of retention.
The more time passes, the greater the difference between traditional learning and SuperMemo.
If you compare SuperMemo with other methods approximating spaced repetition (e.g. priority system, Leitner system, etc.), it may take even longer to demonstrate the difference in retention.
The statement "give it a week before you give up" is to alert a new user to the fact that return on investment in the first week is little. There is a point in time (say several days) before which using SuperMemo may take more time than learning in a traditional way.
Considering the complexity of SuperMemo, difficulty in efficiently formulating knowledge, and the skills needed to master techniques such as "incremental reading", it is important for new users to realize they will need lots of patience before collecting their first ripe fruits of wisdom
(Jerzy Duda, Poland,
Oct 1, 1997)
What is the lowest age at which a child can start using SuperMemo?
The younger the child the more difficult the entry into the learning process. However, with a dose of parental guidance, even first-graders can cope with SuperMemo. The learning process itself is simple and repetitive and the child can quickly enter regular repetitions. Definitely, SuperMemo 98 (and later) at the beginner level is much less daunting than SuperMemo 7 for the initial entry. As a documented example, 9-year-old Agata Czaplinska from Gliwice, Poland, memorized 150 new English words in 2 months working nearly on her own. In another case, 8-year-old Annalynn Clary from Monroe, Louisiana (USA) memorized Cross Country material (1673 items) in 100 days working 30 minutes per day (5 days per week)
SuperMemo contributes to rat race!
(Beata, Nov 11, 2006, 16:59:16)
Your article about Tools : Plan is scary! Don't we get enough of rat race and speed. Your tasklists and incremental reading are not any better. I can see there "more, more, more". And where is the place for our humanity? Did you hear of the Slow movement?
Paradoxically, Tools : Plan can help you slow down. It is up to you how much you pack into the schedule. The main problem with "living slow" is that you may neglect vital matters that will, in time, cause you more stress and hurry than necessary. If you produce a "slow" schedule with a few minutes for handling vital issues on a regular basis, you are less likely to suffer manic "catch up" due to being "slow". Tools : Plan can be abused, but it can also help you reduce stress levels, slow down and yet ... move on at a steady speed.
The same refers to incremental reading. Although it may encourage you to import more and more material, it will painlessly remove the excess import from your view. You can proceed at your favorite healthy speed without ever worrying that you missed anything important. You can import everything, and then optimally read and learn only as much as you got time or patience. The entire process is painless.
Finally, tasklists might belong to the greatest de-stressors of all. Instead of rushing through your to-do list that always grows longer, you can safely focus on top-priority tasks and live with the conviction: "I did not do everything, but at least I did my best".
All in all, tools offered by SuperMemo can be abused and multiply your stress level; however, when used as designed they are supposed to achieve the exact opposite: maximum efficiency at your chosen speed at minimum stress
(William McGhee, Jun
Could SuperMemo be used to extinguish behaviours as well as reinforce them?
Forgetting is a molecular process that cannot easily be induced by natural methods. The more so, there are no sensitive methods to induce selective forgetting, though lesion to some parts of the cerebral cortex may produce roughly localized amnesia. However, there is a component of forgetting that may be influenced. This component is interference. Whenever we learn new things, they always interfere with previously learned material. The interference may enhance some of memories while obliterating others. This fact can be used to employ SuperMemo in forgetting, by formulating and memorizing a large number of contradictory items that strongly interfere with remembered facts that are to be forgotten. For example, if you learn the meaning of the word "indict" and you want to later forget it, you might try to learn words like "indite" or some meaningless like "dictin", "incid", "endict", etc. However, you should not expect the effectiveness of such a procedure to be anything but disappointing.
Which learning method is more effective: traditional SuperMemo with questions and answers, or the new hypermedia SuperMemo with videos, games, puzzles, and tests?
Simple questions and answers are extremely effective and easy to create; however, some users find classic SuperMemo too boring. If the psychological factor plays a part, the variety provided by SuperMemo 8 or later may substantially add to the effectiveness of learning. The answer to the question will depend on the application domain and the mentality of the student.
(Elena and Rachel, US, Dec 12, 1997)
Does SuperMemo improve short-term memory or long-term memory?
SuperMemo builds up long-term memory but helps you increase your mnemonic skills that will result in the impression that your short-term memory works better.
You can also look at this like that: SuperMemo loads knowledge to short-term memory and this is transferred to long-term memory. The effect on long-term memory is stable but the speed of putting things into short-term memory may increase due to training. Short-term memory improvement comes slowly with training, but long-term memory build-up comes immediately upon employing SuperMemo!
Can SuperMemo help remember things that are not stored in SuperMemo?
Yes. To a limited degree. There are three ways in which this is or may be happening:
Delayed repetitions in SuperMemo
(Malcolm Macgregor, Saturday, September 21, 2002 12:32 AM)
What causes delayed repetitions?
Delayed repetitions are usually caused by a user's failure to go through the allocated portion of the learning material. In massive incremental reading, delayed repetitions may also come by design as a result of material overload. Delayed repetitions will always result in an increased measured forgetting index (more forgetting) but can, paradoxically, save student's time by increasing memory consolidation (the speed of learning might be highest for forgetting index around 20-30%). Using Postpone, the student can choose portions of material that will be protected from delay in massive incremental reading. This way, important material takes more time to learn but shows a better recall rate
Trust SuperMemo to save time
(Mike C, Thursday, September 20, 2001 2:46 AM)
I have a problem with Mercy. On 9/10 I answered an item, which was then scheduled again on 9/13. Then on 9/17 I ran Mercy. That item ended up getting re-scheduled to 9/19, today. What I'm seeing is that SuperMemo has assigned it an interval of 9 days rather than 3 days. The problem is that if I answer this item Good or Bright it is going to get scheduled out to something like 13 or 15 days, when it really should be only 4 or 5 days
Once you get a good grade after a longer interval, SuperMemo will naturally use longer intervals as it will upgrade your recall chances. Your anxiety is understandable. However, by shortening intervals you would actually increase time needed for learning. What is worse, artificial shortening of intervals increases the chances of forgetting due to spacing effect
(Colin Quiney, Canada,
Jan 22, 1998)
Do you think SuperMemo can be beneficial in patients with ADD (Attention Deficit Disorder)?
User's ability to focus on repetitions is one of pre-conditions of success with SuperMemo. Seemingly, this would make ADD patients poor SuperMemo learners. Perhaps the report submitted by Maarten Mols from Holland sheds some different light on the issue: SuperMemo in a school for special education.
SuperMemo and Alzheimer's
(Maria Jonas, South Africa, Dec 2000)
Can SuperMemo be used in the treatment of Alzheimer's?
Except for inconclusive anecdotal data, there has been no study of SuperMemo in Alzheimer's. See also: Can too much learning lead to Alzherimer's?
SuperMemo and low self-esteem
(B.B., Friday, October 18, 2002 3:57 AM)
Memory problems? Oh yes! I am sufficiently working at a very low self esteem not only in English, but also in mathematics. Please tell me in some uncertain way what your software does and how it is going to help me in these two fields of study
SuperMemo makes sure you remember what you decide to remember. As long as you show patience and persistence, it guarantees the results. As such, it worked great for people's self-esteem in many cases. However, there are many obstacles on the way. For one, you may dislike the fact that the program seems complex. You may fail to do your daily quota of repetitions. You may fail to formulate your material in a prescribed simple way. As far as English is concerned, you will find it very easy to memorize vocabulary or proverbs or idioms. It will be far harder to learn more elusive things. For example, to make good speeches. Or to recite long poems. Or to debate. Those skills require more understanding of the learning process. As for mathematics, things are even more complex. Without a good manual matching your skills, SuperMemo may appear to be entirely useless. In learning mathematics, the memory is not your main limitation. It is the selection of the material and the way you learn it that will determine your success. In other words, SuperMemo is a sure remedy for memory problems, but it is not a learning panacea. You will find many articles at supermemo.com discussing learning skills such as knowledge formulation, mnemonics, selection of the learning material, etc. See also: Introduction to SuperMemo
Could SuperMemo become an unhealthy addiction?
(smmanic, Thursday, July 07, 2005 12:58 AM)
It seems my entire life is centered on SuperMemo. In making any decision about my life--my choice of job, my choice of transportation means, my choice of social activities--almost invariably I begin with the question, "How will this affect my ability to do repetitions and to add new items?" Clearly I have an obsession with learning and with SuperMemo. My concern, and the subject of my post, is over whether or not this obsession is a healthy one.
In the first six months of using SuperMemo, my focus was strictly on the goal of inputting my total present knowledge (at risk of being forgotten). From a certain perspective, particularly after completing four successful years of college, I am saddened to think I reached that once "monstrous" goal in the space of only six months. Now, after only a year and a half, my collection contains 15,000 memorized items, and my level of knowledge has increased 300%! (That means I feel like I have gone to college three times!) It is a remarkable feeling knowing precisely how much you know.
SuperMemo's instruction manual refers to my database of knowledge as my "collection"; I call it my "brain." My collection is so thorough, it's at least a close replica. With the invention of the thumbdrive, it is humoring to consider that I sometimes wear my brain around my neck!
To be clear, it is difficult to determine whether my present lifestyle is the result of SuperMemo or whether SuperMemo simply fits into a lifestyle I would have chosen for myself anyway. My uncertainty stems partly from the fact that I am a relatively young adult, and like many my age, have never really even begun to settle down into a steady lifestyle and with which I could compare my present "deteriorated" one. (I place that word in quotes because that is how an outsider would likely view it.) Without giving a complete biography of myself, suffice it to say I come from a family whose members all like to move frequently, who tend not to cling too tightly to friends, and whose interests are constantly in flux.
Clearly, SuperMemo is an addiction I cannot shake. The question, again, is whether that addiction is unhealthy and whether I must learn to temper it more. The clichés we throw about would suggest not: knowledge is power and learning what life is all about, right? If the clichés are correct, then anything that helps further my learning is healthy. Perhaps the measure of my obsession's unhealthiness is the extent to which it jeopardizes personal relationships. No, I am not married, but to be fair I have always been somewhat averse to the idea of marriage. I have friends, though my friends have always numbered few. And those relationships I do retain are enhanced, I think, by the increase in confidence and self-esteem attributable to SuperMemo. Despite whatever balance I have managed to maintain, however, I think I would give much of it up if only by doing so could I continue to use SuperMemo. To be sure, knowing now that the program exists, I think if ever I were to lose access to SuperMemo and my collection, I would become suicidal. That is no joke. God, family, and SuperMemo: that is the order of my priorities.
Am I headed for disaster?
You might be heading for disaster, but you might also be heading for greatness. It all depends how well you rationalize your attitude and learn to employ new knowledge in accomplishing your goals.
First you might want to figure out if you are indeed dealing with an addiction. Some researchers argue: "No harm, no addiction". However, there might be a neurophysiologic process that may lead to harm in the future. If you were to, as per your own words, become suicidal at withdrawal, you would indeed meet a key criterion of addiction, but your description seems to indicate that you are rather dealing with a variant of a compulsive behavior. You do not mention tolerance (diminishing satisfaction), lack of control, or desire to reduce the addictive behavior (i.e. conflicting rewarding and punishing stimuli). Interestingly, you are definitely not prone to hide your addiction from the world or lie about it. The fact that you posed the question on a public forum might even indicate that, to a degree, you are proud of your attitude (e.g. while mentioning "doing college in 6 months").
Few people realize how powerful a role genes play in behavior. Thus you might be surprised with a claim that your compulsion is quite likely to have a genetic background. You will certainly notice that none of your friends is prone to acquire a similar habit, and that you will find it hard to locate the roots of the habit in your childhood or upbringing (except where personality traits are amplified, e.g. by inspirational tutoring). Compulsive behaviors do run in families. In this context it would be interesting to study your predisposition to other addictions (as these often correlate). However, this should not be a reason to worry per se. Just the opposite. The history of invention and scientific discovery proves that those traits might be responsible for quite a portion of mankind's progress. In the end, your compulsion may be injurious to your own life, and, at the same time, bring immense benefits to others. This has often been the outcome in the life of great inventors or scientists who changed the world while living in utter misery. Unlike animals, humans seem to have strong anti-entropy mechanisms imprinted in their brains. In other words, they tend to marvel over art, music, patterns, structures, models, abstract concepts, beauty of science, etc. Compulsive behaviors may be a strong expression of these mechanisms. You love the fact of being in control over what you learn and that must be rooted in the beauty of building a neat structure of knowledge. The "rage to learn" is one of primary characteristics of gifted children. If you now employ this compulsion to accomplishing lofty goals, greatness is likely to outrun disaster as the likely outcome. In other words, you must resolutely plan against learning for art stake. Instead, your learning, targeted at specific goals, should make up a rational proportion of your time with the rest filled out largely by creative activities targeted at specific productive goals. In Stephen Covey's terminology, you should balance your production to production capacity ratio (P/PC). If you accept well-selected knowledge as valuable you should be able to avoid conflicting stimuli and an internal battle of conscience.
This is an interesting material for more than just an FAQ. It would be nice if you stayed in touch to develop this into a case study.
SuperMemo will not help you bring things to mind without a trigger
(D.M., Jan 26, 2006, 06:54:40)
How can I use SuperMemo to remember to remember something?
SuperMemo will not help you much in bringing things to your mind without a behavioral trigger. It only helps you remember the association between the stimulus and the desired response. It will not help you generate a response without a stimulus. You cannot use it to remember to turn off the gas, unless you associate the turning off the gas with a specific trigger. For example:
Q: What should I always remember about at the time of leaving
A: Check if the gas is off
With some training, checking the gas will become a habit. SuperMemo will only ensure that you do not forget about forming the habit
I don't buy memory overload hypothesis in Alzheimer's
(Mike, Tue, May 28, 2002 5:01)
I'm sure the basic premise of the "memory overload hypothesis" is flawed. The effect of SuperMemo is really no different, at the basic biochemical level, than any other repetitive learning activity (although certainly more efficient). If I dial my mother's phone number 100 times over the next few years, does that also contribute to Alzheimer's?
It is important to differentiate between recalling/reusing old memories, and forming new memories. When you redial the same phone number, you are not likely to form new long-term memories. Nor does the overall storage change when you make repetitions of outstanding items in SuperMemo. However, the power of SuperMemo is in that the efficient rehearsal of the old material provides more time for learning new material. Consequently, you will be forming far more new memories than an average non-SuperMemo student. This would be exactly what Robin P. Clarke worries about: too much new stuff in your head. However, the evidence leaves little room for truly harmful overload (e.g. as in cramming in stressful conditions). If the hypothesis was right for the subset of cases violating the principles of mental hygiene, well-employed SuperMemo would not classify
It is possible to roughly estimate the amount of time needed for learning a portion of material
(Patrik Nilsson, Tuesday, December 18, 2001 11:53 AM)
Can Tools : Statistics : Simulation help me figure out the date when it is expected that I manage my learning material?
Once you introduce an item into the learning process and execute all outstanding repetitions, it is SuperMemo's responsibility to ensure the desired level of retention. For that reason, you can assume that you "managed" your material as soon as it has been introduced into the learning process. Consequently, you do not need to run Simulation. If you want to memorize 1000 items and you decide to introduce 10 items per day into the learning process, you will "manage" the material in 1000/10=100 days. Now you can use Simulation to try to estimate how much work this will require. A rule of thumb is that you need 10x more effort for repetitions than for learning new material. This could indicate that if you memorize 10 items per day, you may expect 100 repetitions per day (at least in the initial period). This number may vary greatly depending on the difficulty of the material and your learning skills and techniques. Use Simulation to get a better estimate. With rescheduling tools (e.g. Postpone), you can also reduce the daily load of repetitions; however, you will then suffer some loss in retention. Another rule of thumb is that to increase your forgetting index from 10% to 20% you would need to either (1) massively overload the learning process (e.g. by increasing the inflow of material 10-fold), or (2) dramatically cut down on the learning effort (e.g. by 90%). For more see: Theoretical aspects of SuperMemo
Vacation may feel like "improving memory"
(p.b., Thu, Apr 11, 2002 18:02)
When I came back from vacation, where I did not do my daily SuperMemo, some 1,500 items awaited me. I found out with surprise that I recall the items better after this 10-day break then on usual days with daily repetitions. Maybe my mind had so good rest that it worked noticeably better?
If the vacation took away some stressload or gave you a chance to catch up with sleep, your fresher mind might partly explain a higher recall rate. However, you should also be aware of the placebo effect enhanced by the surprise that you remember so much. If your forgetting is usually 10%, you might have expected 30% after the vacation (as most people do expect). When you had noticed or sensed it is 12% (the more likely value), you could have felt as if it was 8%. It would "feel" as if your scores improved even though they were worse in proportion to the length of the delay
Is repetition the best way to remember?
(BruceTrek, Sunday, March 10, 2002 7:20 AM)
Can you send me some links to independent research that shows repetition is the best way to remember?
There isn't much research being done currently to show that repetition is the best way to remember. This fact is just taken as true. Similarly there isn't much research done to prove that the brain is the organ where thoughts are born. In the latter case, nearly all neuroscience lives by the brain-mind connection. This does not prevent others (beyond science) to look for mystical explanations to the thought process, emotion, consciousness, soul, etc. As for the repetition, no reputable textbook on the learning methods will deny that "repetitio mater studiorum est". Some researchers in the field looked to confirm the concept of permastore (i.e. lifelong memories), but that effort has not been successful. Some "accelerated learning" companies will try to sell you learning materials and insist that "it is all in mnemonic representation", i.e. you can learn it once (the "right way") and remember it for ever. The best way to disprove that claim is to learn mnemonic techniques, buy those "optimally formulated" materials and learn it with SuperMemo. SuperMemo will measure and prove that well formulated material is indeed remembered much better. But it is still subject to the same old negatively exponential forgetting process that affects all acquired memories (as opposed to innate memories wired into the brain at birth). Today, supermemo.com is probably the only website so heavily focused on the concept of repetition. You will find links to many sites throughout but there isn't one that would particularly attempt to prove that repetition is the key to lasting memories. For every user of SuperMemo, the question on the importance of repetition becomes irrelevant after a few weeks with the program. SuperMemo measures the forgetting rates and graphically illustrates what happens with memories once they are not taken care of
Minimum information principle v.s. the length of "20 rules" article
(NF Lynch (Isr), Israel, Wednesday, January 15, 2003 9:29 AM)
Your list of The 20 rules of formulating knowledge in learning flies in the face of your own learning principles. The list is too long and without any hierarchical basis or mnemonic hooks
The minimum information principle refers only to active knowledge. Passive knowledge can be processed incrementally. In incremental reading, there is no upper limit on the length of texts that can be processed efficiently. Once the individual rules listed in the article are memorized using, for example, cloze deletions, they need to meticulously adhere to the minimum information principle. In other words, articles you read can be outrageously long. It is the answers to questions you want to remember that must be short
There is no harm in "memorizing" things you already know
(Comcast Mail , Saturday, December 20, 2003 7:11 AM)
How do I add to SuperMemo knowledge that I previously acquired (i.e., facts that I already know or that I only recently forgot but initially learned without the aid of SuperMemo)? I want to add this knowledge to avoid forgetting it again
You can add this knowledge in the same way as facts that you do not know. There is little harm to the overall learning process as this material will quickly be classified as "easy". To save time on needless repetition, you could additionally set the first interval on individual items (or topics) to an appropriately longer value (Ctrl+J). For example, if you believe you are not likely to forget an item for 3 months, you could set the first interval to 99 days or so.
I used SuperMemo 2 shareware, and was accustomed to repeating forgotten items on the next day. It is very irritating that in SuperMemo for Windows I do not have this possibility
SuperMemo will schedule forgotten items in intervals that are determined by the forgetting index. The greatest increase in the speed of learning in newer versions of SuperMemo as compared with SuperMemo 2 resulted from substantially increasing the length of the first interval. The student may be left with the feeling that he is likely to forget the item again if it is not repeated on the next day. Statistically, however, he will forget no more than the proportion defined by the forgetting index (specified in Tools : Options : Learning : Forgetting index). By reducing the forgetting index to less than 5%, the length of the first interval is likely to drop to 1-2 days in most cases. Moreover, if you are particular about repeating a given item on the next day, you can choose Ctrl+M to commit or recommit an item with a selected first interval
Poland, June 3, 1997)
Personal question to Dr Wozniak: If you did not have SuperMemo at hand, which competitive product would you use?
I would probably pick one of the programs that most closely follow the SuperMemo paradigm, e.g. Edukom or PowerMemo. Only later would I choose from better known products that are less focused on spaced repetition such as Langmaster or YDP's Collins Dictionary. However, as a software developer involved in the design and implementation of SuperMemo since 1987, I would rather opt for implementing the program from scratch. Back in 1987, the first version was written in 16 evenings. I would have to give up all the bells and whistles, but it is the core that matters most. And it would be most difficult to give up the control over what new options might yet be implemented
Intervals are measured in circadian cycles
(Michael M., United Kingdom, Friday, March 04, 2005 10:47 AM)
It would be a nice option to have intervals calculated in minutes, so that if one first learns an item at say 08:30, it might be re-presented at say 12:15, then 19:45, then the next day at 10:45 etc.
Due to the spacing effect, measuring intervals in minutes might be counterproductive. Not only would extra necessary repetitions be needed, but also the spacing effect might be evoked with the resultant decrease in memory stability. The main reason for rounding intervals to days is the memory consolidation cycle measured by waking-sleep rhythms. Although it is recommended to make repetitions at times of peak mental performance, there is little long-term difference between making repetitions in the morning and making them in the evening. In particular, this difference is negligible if you choose one of the two mental peaks of the day: early morning and early evening (the terms "morning" and "evening" are relative and an owl's peak may come at 2 am). If you believe that your items require repetitions that are more frequent than 24 hours, be sure they comply with the rules of formulating knowledge
Advanced repetitions and the spacing effect
(Malcolm Macgregor, Saturday, September 21, 2002 12:32 AM)
Could you explain how repetitions advanced in time relate to the spacing effect?
Before an important exam, the student may decide to review a portion of material even though individual items may be scheduled for repetition months or years into the future. The spacing effect prevents memory consolidation when inter-repetition intervals are far shorter than the chosen optimum. For that reason, formulas used in earlier versions of SuperMemo cannot be used to compute a new interval after the forced premature repetition. Instead, a heuristic formula is used to shorten the interval accordingly. For example, if the optimum interval was 1000 days and the next optimum interval was to be, say, 1300 days, SuperMemo will use the interval range from 1000 days to 1300 days for premature repetitions. If the repetition takes place 10 days after the last review (instead of 1000 days), it will have nearly no effect on memory consolidation and SuperMemo will choose 1000 (or 1001) as the next interval. However, if the repetition takes place on 900th day, SuperMemo will assume it is nearly as effective as the optimally scheduled repetition and schedule the next repetition in, say, 1297 days. The formula for "interval attenuation" is not based on hard science but is good enough for practical applications where advanced review is an exception rather than a norm
Learning to do calculations in memory
(Dragan Vidovic, Netherlands, Wednesday, March 10, 2004 9:25 AM)
I am trying to learn to calculate quickly in mind. I am a mathematician, and this is a hard problem form me, because I use a calculator since I was ten. The main problem with calculating in mind is memory
You are right. The amazing feats of Rudiger Gamm are based on memory and the right techniques. You should begin with learning mnemonic techniques relevant to mathemetical calculations (e.g. peg system). Then you could try to learn a couple of rudimentary SuperMemo collections (e.g. Multiplication 20x20, Division 20:20, Division 1/20, etc.). While working with basic collections you will hone your own mnemonic techniques useful in basic calculations. Then you can gradually expand the difficulty of the material used in your practise. Ultimately, if you aim high, untold hours of practice are sine qua non of success (Gamm used to spend whole days practising).
One of 1985 experiments was inconclusive
(Tomasz P. Szynalski, Saturday, August 04, 2001 1:33 PM)
I see that one of early experiments showed that constant 18-day intervals were better than increasing intervals. So how does this prove the validity of SuperMemo?
This 1985 experiment does not prove validity of SuperMemo. Nor does it disprove it. In that respect, it is simply inconclusive. The experiment was an attempt to "guess" optimum increase in intervals and the guess appeared wrong. If the underlying hypothesis was that increasing intervals are always better than equally distributed intervals, then this experiment proved that hypothesis wrong. Obviously, if equal intervals are short enough or the increase in the length of intervals is too fast, equal-interval schedule may appear superior. This is why, in the long run, we need to use a computer to optimally adjust the repetition schedule to the defined level of knowledge retention
Spaced Application & Semantic Networks
(Mark G. P., Thu, Jun 21, 2001 6:24)
I read a question in Medical Biology:
Q: mem: What is the impact of blocking phospholipase A2 on inducing LTP?
Uhhhhhhhh …… but if I don't have a clue what phospholipase A2 is then this item becomes meaningless. Sure, if I pound away at recalling fuzzy, shallow memories of meaningless items, eventually I may encounter the information I need and the lights will come on, but is this the most efficient way to learn?
Rule #1 in selecting and formulating your learning materials is: Do not learn what you do not understand! (see: 20 rules...).
The correct action upon encountering such an item is one of the following:
Ideally, you should create your collections on your own; however, ready-made material may also be useful if it belongs to a well-defined and well-targeted subset, or if it is used as a complementary inspiration rather than a supplement for wider study. Medical Biology is sorted by difficulty but it is unlikely to be fully understood even by students of medicine. For that reason, complementary research and incremental reading will always make a vital addition to studying this material
Use Simulation to figure out the expected speed of learning
(Robyn, Sep 23, 2004, 12:47:42)
If an individual can commit to exactly 100 repetitions per day, how large will his collection eventually become?
You can use Tools : Statistics : Simulation to answer similar questions. For 100 repetitions per day, you will on average arrive at 75,000 items in 15 years. However, this number will vary greatly depending on the difficulty of your material, your mnemonic skills, forgetting index, the use of postpones, etc. If you do not have SuperMemo, you can roughly estimate the number of items by assuming that you can learn 1 new item per 7-10 items repeated per day. Using this method you will expect 55,000-78,000 items memorized in 15 years. Remember, however, that the learning speed is not exactly linear and you will get better results for shorter periods and slightly worse results for longer periods. For more see: Theory
Memorizing instances of abstract cases is not a waste of time
(SuperMemo R&D, Wed, Aug 01, 2001 21:47)
In SuperMemo you propose to introduce several examples of the same rule. For example, 13*10=130, 24*10=240 and 69*10=690. One could get dozens of different possible appearances of one principle. Does this make sense?
Yes. This is instance training. Problems you solve are instances of a certain abstraction. A single rule-based item requires a very smart mind to produce a strongly applicable abstraction. Several instance items help you make use of the natural properties of neural networks to enforce the abstraction. In many cases, you will achieve better results by memorizing a rule and several instances of its application than by just the rule alone. In SuperMemo, the cost of such a redundancy is negligible. Remember that in SuperMemo you spend most of your time on repetitions of difficult material. Redundancy improves retention by optimizing representation. Paradoxically, by adding redundant instance items, you can often reduce overall workload. And even if the workload were to increase, the applicability of thus-strengthed abstraction is a highly welcome side effect
Limited Postpone will not damage your learning process
(Mariusz, Jun 17, 2002)
How much does my retention drop when I use Postpone?
For an ideally exponential forgetting curve we can show that:
The table below shows that even a large number of postpones is insignificant as long as the delay factor d is low (you can choose this number in the postpone dialog box). However, the default delay factor of 1.1 will increase the forgetting index from 10% to 20% if postpone is executed successfully seven times in succession.
Table: Increase in the measured forgetting index as a result of using Postpone among items with the requested forgetting index equal to 10%. Grey column headers: delay factor. Yellow row headers: number of postpones. Table body: measured forgetting index. Blue area: measured forgetting index above 20%
|p / d||1.01||1.02||1.03||1.05||1.1||1.2||1.3|
First meaningful split into difficulty categories occurs at first repetition
(Aaron Koller, USA Educational, Thu, Jun 13, 2002 5:25)
Ten days ago I had a collection with about 200 items in it. I reset the collection and learned all 200 items that day. Starting the next day, I added about 20-25 more items each day. My problem is: SuperMemo is scheduling the next repetition for the newly added items up to 5 weeks in the future, seemingly without regard what score I give it. It seems I will certainly forget an item I didn't know today if I won't seen it again until next month
It takes time for SuperMemo to be highly accurate with ensuring the requested retention of knowledge. Your poor grades next month will help it improve its performance. The first interval of all new items is computed with the same forgetting curve. This means that both difficult and easy items will be scheduled at the same time. The first grade given after this uniform interval is the first tool SuperMemo uses to estimate item difficulty. It cannot use grades provided while memorizing because these can pertain to completely unknown material as well as to material that has just been introduced manually. If you scored excellent on those first 200 items, all following items will be first scheduled at a much longer interval (i.e. the optimum interval for new items whose difficulty is not known). In simple terms, SuperMemo has no way of knowing that you have already worked with the first 200 items and your grades are exceedingly good. Nor can it know that the follow-up items are far more difficult. The only hint of irregularity will come with grades received in repetitions. For that reason you will need to wait as long as the first interval (up to 20 days) and you will need to process comparably many difficult items (i.e. around 200) to proportionally shorten the first interval. The average first interval may reach 20 days, but rarely, due to interval dispersion you might see items scheduled in over four weeks (i.e. 20 days plus the dispersion interval which follows a normal distribution).
You have two choices:
A good practice is to keep all your knowledge in one collection to avoid similar deviant behavior. All forms of intermittent learning (e.g. relearning the material, breaks in learning, etc.) make it difficult for SuperMemo to build a good memory model
The optimum number of repetitions will depend on numerous factors
(BruceTrek, Sunday, March 10, 2002 7:20 AM)
Is there any research or general knowledge about the ideal number of repetitions to use when you learn something new?
There is no ideal number of repetitions. Depending on the student and the complexity of knowledge, the optimum number of repetitions may vary from a few to dozens in lifetime. For ill-structured knowledge, achieving retention of 90% may be practically impossible. The best research tool here is SuperMemo. Its algorithm will attempt to predict how a single piece of knowledge in your collection will fare. You can peek at Future repetitions field to see how many repetitions are statistically expected within the next 30 years for the selected level of knowledge retention for that particular piece of knowledge
SuperMemo should not be viewed as a cramming tool
(Gundam Fool, Monday, February 18, 2002 1:00 AM)
What about applicability? For example, I memorize a math formula with SuperMemo, should I also add an item that makes me use this math formula so I also remember how to apply this math formula to a problem? Or is the associative thinking part separate from SuperMemo and I should only memorize facts and formulas?
You can use SuperMemo to boost creative thinking as well. The optimum approach to formulas will depend on your particular needs. Simple formulas may be self-containing and applicable in many contexts without additional clues. The connection between others and relevant models may be less clear. Then, depending on your needs and the priority of the formula, you can add examples, anecdotes, loose associations, mutated formulas, formula derivation, instances to solve, etc. If you do not have much experience with SuperMemo, start with raw formulas and see how they work in practice. If a formula becomes "dry" and unassociative, you can support it with additional material. You can, for example, think over the formula and its known association and describe it in prose in your own words. Then you can dismantle that prose in incremental reading and gradually build a better understanding of the formula's applicability. Using prose and incremental reading will be equivalent to perpetuating your today's understanding of the formula and, as a bonus, deepening associations with individual components of the model
SuperMemo does not use microspacing of repetitions
(mgpatterson, Thursday, August 23, 2001 5:16 AM)
While I have complete confidence in SuperMemo scheduling algorithms, I needed to modify the operation of the final drill. From my research I've concluded that it is critical during initial encoding (when a student is first learning an item) that they are required to recall it at least 3 times during a 30-minute window--even though they may fully "know" the item each time. This repeated micro-spaced active recall (overlearning) leads to a neurochemical cascade that shifts the association into Intermediate-Term memory (lasting up to 3 days)
Unfortunately, we cannot agree that microspacing is an effective way of using student's time. We are not aware of neurochemical variables needed to consolidate long-term memory other than those described in the SuperMemo model. In this light, final drill is only needed as a tool for coping with negligent/imperfect repetitions. Ideally, a single repetition and activation of short-term memory is all that is needed to begin the long-term consolidation phase. We are open to discussing your research in detail though
Forgetting curve for ill-formulated items is flattened
(anonymous, Wednesday, July 25, 2001 2:54 PM)
I had a 5-months break in using SuperMemo. I resumed my repetitions and noticed that I still remembered many items. Initially, SuperMemo asked me to repeat difficult items (e.g. enumerations). To my surprise, I remembered many of these items. SuperMemo required a 15 days interval, while I made my repetitions after 150 days and still succeeded. I no longer believe in the optimality of the SuperMemo algorithm. Probably it is 10 times worse than optimal
Your results are in full compliance with theoretical expectations. It is no surprise that SuperMemo initially tossed the most difficult items at you, and it is no surprise that you showed remarkable recall on these items. Those items clearly belong to those that have not been formulated in compliance with representation rules (e.g. enumerations are notoriously difficult). If you imagine memories as sets of apples (you can see an apple as a single synapse in the brain), good memories are like small collections of well-polished apples. Bad memories (e.g. enumerations, complex items, ambiguous items, etc.) are like large collections of apples of which few are spoilt. Naturally, spoilt apples rot fast and make recall difficult. After just 15 days, all spoilt apples might have been rotten. During the remaining 150 days, the remaining apples rot very slowly. Hence the typical departure of wrongly formulated items from the shape of the classical forgetting curve. For bad items, the curve is flattened (as an expected superposition of several Ebbinghausian curves). SuperMemo blindly obeys your recall criteria. If it takes 15 days to ensure 98% recall, SuperMemo will take no consideration of the fact that at 150 days you may still show 95% recall. This is why SuperMemo 2000 or later includes leech management. It makes it easy to identify bad items and use auto-postpone or auto-forget options. Auto-postpone will do exactly what you expect, i.e. delay bad items with little impact on overall retention. Auto-forget will help you rebuild memories from scratch. Occasionally, the newly established memory representation will click and your recall will improve. Naturally, the best method against bad items is the use of appropriate representation (see: 20 rules of formulating knowledge for learning). Interestingly, SuperMemo can never predict the moment of forgetting of a single item. Forgetting is a stochastic process and can only operate on averages. A frequently propagated fallacy about SuperMemo is that it predicts the exact moment of forgetting: this is not true, and this is not possible. What SuperMemo does is a search for intervals at which items of given difficulty are likely to show a given probability of forgetting (e.g. 5%). If you look for a numerical measure of the algorithm's inaccuracy, instead of comparing intervals you should rather compare retention levels as the retention is the optimization criterion here. Even for a pure negatively exponential forgetting curve, a 10-fold deviation in interval estimation will result in R2=exp(10*ln(R1)) difference in retention. This is equivalent to a drop from 98% to 81%. For a flattened forgetting curve typical of badly-formulated items, this drop may be as little as 98%->95%
Flatter forgetting curve does not increase optimum interval
(Tomasz P. Szynalski, Saturday, August 04, 2001 5:53 AM)
If the forgetting curve is flatter for difficult items, I will remember them for a longer time, right? Does that suggest that ill-formulated items are remembered better?
No. Flattened forgetting curve will increase retention measurements in intervals that are a multiple of the optimum interval as compared with the typical negatively exponential curve for well-structured material. However, the optimum intervals for ill-formulated items will expectedly be shorter as can be observed on the first interval graph in Tools : Statistics : Analysis : Graphs : First Interval. The smoothness of this graph depends on the number of repetitions recorded. In the picture below, over 90,000 repetitions have been recorded
There is no remedy to interference
(Justin, Wed, Nov 07, 2001 23:58)
What should I do when I get two successive elements that are identical but flipped (i.e. the question goes in the place of the answer)? The first card obviously "refreshes" my memory -- so the "testing" of the second item is inaccurate. I usually just click Fail ('D'), which is bad for the learning process
Your only remedy is introspection and an honest attempt to estimate how you would have scored had the refreshing item not been presented. Otherwise, scoring Bad or Fail should not do much damage to the learning process. In the long run (i.e. when intervals increase) and with a high volume of the learning material, this is usually not a problem to worry about. Successive occurrence of similar items is rare in such circumstances
20 rules of formulating knowledge and SuperMemo Decalog are springboards for learning with SuperMemo
(Spud Science, USA, Feb 14, 1998)
What is the best (most effective) way to set up a new series of question and answer pairs in SuperMemo?
See: 20 rules of formulating knowledge. See also: SuperMemo Decalog. For a more academic reading in the subject you can have a look at Knowledge Structuring and Representation...
Tackling contextual dependency during repetitions
(Jake White, USA, May 14, 1997)
Should not final drill continue keeping a queue of no less than 10 newly learned words in order to make sure that when repeated again and again they will really be imprinted well in short-term memory.
This solution may indeed eliminate contextual dependency in final drill (remembering items only because of having them in a given context); however, this would involve lots of extra repetitions that would contradict the principle of SuperMemo: maximum effect at minimum time. Additionally, the learner would risk activating the spacing effect, i.e. reducing the probability of recall as a result of excessive repetition! The best solution to contextual dependency is:
For example, import your mind map as graphics to an image
component and check Answer on the image component
menu. Add a text component, e.g. "What is the structure of mind map X?".
During repetitions grade yourself less than Pass (3) each time your forget any part of the mind map! Do not forget that you should reinforce "weak links" in the mind map with separately formulated items of simple question-answer or question-picture form. Each time you forget part of the mind map, see if you have reinforced the forgotten links in separate items!
No specialist knowledge is needed to prepare simple material collection in SuperMemo
(Anatolyi Lipatov, Ukraine, Jul 12, 1998)
I am using Advanced English to enhance my English and business English. Now I am registering for CFA examination (that is Chartered Financial Analyst program of Association of Investment Management and Research). There are several organizations developing and distributing methodological stuff for preparing to the exam. A lot of things should be memorized for passing the exam. What do you think the best way to fit SuperMemo for memorization is and what approach should I use to prepare my own knowledge base for memorizing the material. Is special programming knowledge needed for it?
No specialist knowledge is needed to prepare simple learning material collection in SuperMemo. With Alt+A (i.e. Add new) you get the core functionality! All advanced editing options can be worked around by an appropriate questions-and-answer approach. Perhaps it would be useful yet to learn how to add images to your items (see help for details). To learn more about effectively structuring knowledge in SuperMemo you might want to read 20 Rules of Formulating Knowledge and Knowledge Structuring and Representation; however, nothing works better as learning on one's own mistakes in formulating knowledge for learning with SuperMemo
Forgetting index vs. retention
(Tomasz Szynalski, Poland, Oct 18, 1998)
What value of the forgetting index ensures the optimum ratio of (retention)/(time spent per day)?
Paradoxically, the highest speed of learning can be accomplished ... without SuperMemo! In our daily life we pick up lots of facts that stay in our memory for long with few repetitions in lifetime! The problem is that these are usually not exactly the facts or rules that are critical to our goals. In other words, not the speed of acquiring new items counts but the speed of acquiring new items bearing a given content.
It is difficult to determine exactly what forgetting index brings the highest acquisition rate. Simulation experiments have consistently pointed to the value of 25-30%. You can even plot speed-vs.-forgetting graph using your own actual learning material in SuperMemo using Tools : Statistics : Simulation. You will probably also arrive to similar results
As you perhaps know, SuperMemo disallows of the forgetting index above 20%. This comes from the fact that you should aim at achieving high speed of learning combined with high retention of the learned material. Setting the forgetting index above 20% would be like giving up SuperMemo altogether and coming back to remembering only that what is easy to remember. In highly interlinked material where new knowledge depends on the previously acquired knowledge, high forgetting rate can even be more harmful
Nevertheless, if you want to maximize the speed of learning with little control over what actually stays in your memory, set the forgetting index to 20%
Breaking one long repetition session into a few smaller ones is generally beneficial
(Robert Szumilo, Poland, Jan 3, 1999)
What is the optimum approach to making repetitions with SuperMemo: one long session or a few smaller sessions (e.g. main repetitions in the morning and the final drill in the evening)?
For psychological reasons, the quality of learning should increase substantially when working in separate sessions, esp. if the number of repetitions surpasses 100 per day. Additionally, a break before final drill is useful due to the spacing effect. The danger of this approach is ... you can easily drive yourself into a situation in which you will spend excessive proportion of your day on repetitions (in the future when your schedule changes you might have problems with keeping up with your present pace)
Five-day exposure is not enough to retain 65% of knowledge for 15 years to life
(Tony D'Angelo, USA, Feb 3, 1999)
I am a management consultant who uses professional development programs from a company called Resource Associates Corporation. These programs are based upon spaced repetition learning. In their literature they cite an unnamed study that suggests that people will generally only retain 2% of information they are exposed to in a one time event after 2 weeks. In contrast they suggest that at least 65% of information delivered over 5 consecutive days may be retained for 15 years to life. Can you confirm those claims?
If you look at the graphs generated by SuperMemo during the learning process, you will notice that the 2% figure might be true depending on how the material has been presented and its difficulty. However, the claim that 65% of knowledge can be retained for 15 years as a result of five-day exposure must be false in the same light. Even if we consider perfectly formulated knowledge (i.e. knowledge characterized by the highest possible A-Factor), you might need to space the quoted five exposures in the period of 2-3 years to make the 65% figure realistic. If the quoted figures were accurate, you would probably never need to use SuperMemo! For more details on the speed of learning, see Theoretical aspects of SuperMemo
Your repetition schedule in SuperMemo will never become overbooked
(Lawrence A. MacDonald, USA, Feb 16, 1999)
What is the result on scheduling in the future when you add a question every day? At what point does your future schedule become vastly over booked?
The beautiful thing about SuperMemo is that overbooking never happens! Your question and worry is typical for those who begin their work with SuperMemo. Please have a look at simulation experiments that show that the learning curve, in the long perspective, is nearly linear, i.e. it does not bend as a result of "overbooking"! In practice, you will be able to notice the decline in the speed of learning for no more than a year. Later the slowdown is minimum and entirely imperceptible! The mathematical explanation of this is quite complex. The measurements show that you spend 50% of time on just 2.5% of the material! Imagine then that the memorized material quickly reaches high intervals and disappears from view. You just constantly struggle with newly memorized items and items that you find difficult to remember
SuperMemo will likely help you strengthen your mnemonic skills
(Tomasz Szkopek, Poland, Feb 25, 1999)
Does using SuperMemo increase the capacity of the human brain for learning with traditional methods?
SuperMemo is likely to strengthen your mnemonic capacity. It is does not include any specific options for that purpose, it simply acts as a training tool via repetition. Your mnemonic techniques will develop subconsciously as a result of training the same way like muscles grow as a result of lifting weights. However, this should not be treated as a substitute for a course in mnemonic techniques.
Secondly, recent research shows that the brain continues growing new neurons and new synapses even at a very advanced age. This process is dramatically enhanced by rich environment and intellectual stimulation. It is possible that a 40-minute session with SuperMemo also acts as a good stimulant of these growth and development processes (as much as any challenging intellectual activity)
Postpone is a tool for those unable to spend time on learning on certain days
(M.R.W., Poland, March 24, 2000)
SuperMemo does not seem to provide enough support for those who cannot afford to spend their time daily on regular repetitions. The daily regimen may discourage a large number of potential users. Sometimes one would like to double the effort. On other occasions, one might not feel like working with the program at all. For example, I could never convince my father to use SuperMemo. When he comes back from work, he is often too tired to watch TV, let alone make repetitions. Do you plan to include special tools that would make it possible to learn only on weekends or exclude certain days of the week?
Weekly calendar of repetitions is in consideration; however, this option adds complexity without actually providing much learning benefit. It may actually appear harmful:
How can I review my material on the same day I add it to SuperMemo?
(NamJongmin, , Monday, October 12, 2009 2:35 PM)
Why is that I can't review the material right away, but have to wait a couple of days. I know that's to achieve the benefit of spacing effect. Yet, is there a way to review the material right away?
We do not review the material on the same day because of the spacing effect that will make the review ineffective or even counterproductive. We also do not do it to have more time for learning new material, or for doing other things. If you add learning material to SuperMemo today, and you do it with deliberation, reviewing it right away is tantamount to cramming, and you should restort to it only in emergencies (e.g. if your exam comes tomorrow). To review your new items on the next day, open the item set in the browser and choose Learning : Review All. To review your items yet today, choose Tools : Random Test instead. Note that Random Test will leave no trace of the review in your learning process. This means that SuperMemo will not take a correction for whatever memory effect you will manage to produce with instant review.
Medical knowledge: too much!
(German Salazar Pareja, Dec 29, 2009)
You write that "the demands of a medical school go well beyond the human capacity to learn. That's a norm for most schools. Even SuperMemo is powerless here". This is a fallacy, more than one million physicians in the world are your proof. Human capacity to learn is huge. Not infinite, but huge. I don't have the numbers about this capacity, but I could search it if necessary. However, I've also struggled during my med school years, probably because I was concerned with learning the most I could. I believe medical school curricula is designed with other purposes in mind, through you career you'll always have time constrains, the ones who survive this selective process are not only those who study most, but those who learn how to deal the the problem of time constraints and still can do the most of the learning in med school
You need to interpret the quoted sentence in the context of SuperMemo. Basically, in the context of SuperMemo, cramming is not considered a form of learning. Consequently, the capacity to learn, should be understood as the lifetime capacity of the long-term memory storage rather than the ability to store a large volume of information in memory in a short period of time (e.g. before an exam). Similarly, demands of a medical school, should be interpreted as the student-perceived demands on what material is absolutely necessary to know. Those demands are quite different from what material is actually known at graduation (let alone a few years after the graduation). Using this interpretation, if you raise the bar for what you consider "learned", and raise the bar for what you think you need to know, you arrive at the point where it is not possible to learn all that you think you should know as a medical graduate. There are indeed millions of active physicians today, however, each would fail some basic medical knowlege questions when tested today. This is not to say that all physicials are bad physicians. This is only to say that each and every human being, however smart, has his or her own patches of ignorance, even within his or her own field of expertise.
Two methods can be used to easily illustrate how university knowledge exceeds human capacity to learn:
The term "staggered learning" may sometimes be used to describe "spaced learning"
(SKlein, Holland, Tuesday, December 05, 2000 7:32 AM)
What is the difference between staggered learning and learning based on spaced repetition?
The term "staggered learning" is not used often. It may refer to intermittent learning or learning based on repetition and review. It is used in reference to the curriculum rather than the method of learning. Probably, its association with spaced repetition comes from the fact that it was used in the context of the Leitner method which is a very old form of spaced repetition